Despite the fact that Medicare spends millions of dollars annually on chemotherapy for elderly cancer patients, surprisingly little is known about the extent to which cancer chemotherapies help or harm such elderly patients. This unsettling paradox is the direct result of the well-described under-enrollment of elderly on the clinical trials of chemotherapy. Through this empirical research that integrates both clinical trial and observational data sources and methods, we seek to narrow this critical gap in clinical knowledge by studying the survival outcomes of both unselected elderly Medicare beneficiaries and their carefully chosen clinical trial counterparts. In Aim 1, we describe survival for stage-specific cohorts of elderly Medicare patients with breast, colorectal, lung, and pancreas cancer (n=100,000) following first-line treatment with one of several standard chemotherapies. In Aim 2, we describe the survival of patients from Aim 1 relative to untreated, but otherwise matched similar elderly Medicare patients with cancer (n=100,000) using propensity score methods. In Aim 3, we compare attributes of the cohorts of population elderly Medicare patients who were treated in Aim 1 to those of similarly treated clinical trial elderly Medicare patients. To do this, we create a new data set termed the CALGB-CMS data that relies on linkage (at the individual patient level) of Cancer and Leukemia Group B (CALGB) clinical trial data pertaining to elderly trial enrollees (n=4,000), to their contemporaneous observational Medicare data and other extant administrative data sources in a manner that parallels the SEER- Medicare data structure. After appending SEER-Medicare observational data to the CALGB-CMS data, we compare SEER-Medicare observational and CALGB-CMS clinical trial patients according to attributes of patients (e.g., demographics, comorbidity) and providers (e.g., board certification, years in practice). In Aim 4, we compare the survival outcomes of the population-treated elderly SEER-Medicare patients to those of the clinical trial-treated elderly CALGB-CMS patients following receipt of the same standard chemotherapy regimens. We estimate survival following therapy according to patient type and use multi-level approaches to identify and quantify salient patient and provider determinants of any observed differences in survival outcomes between population-treated and clinical trial-treated patients. In Aim 5, we create clinical prediction models for physicians to estimate survival for individual patients under treatment and no treatment scenarios; the models will tailor estimates to additional patient factors like age, sex, race and comorbidity. Through the proposed research, we will better understand the expected survival of population-treated elderly Medicare patients following the standard chemotherapy regimens whose efficacies were established in clinical trial patients. The work will help practicing oncologists in their care of elderly patients and set the stage for further study of the survival benefits of chemotherapy in the elderly. PUBLIC HEALTH RELEVANCE: While more than two thirds of all cancer patients in the United States are diagnosed at or after age 65, less than one third of participants in chemotherapy clinical trials are 65 or older. This well known under-representation of the elderly on chemotherapy trials has created an enormous void in clinical knowledge regarding the risks and benefits of even standard chemotherapy regimens in the elderly; this void may compromise informed decision-making by physicians and their elderly cancer patients and result in inappropriate over and under-treatment of the elderly. This research is focused on studying the survival outcomes of elderly Medicare cancer patients following treatment with what can be considered standard chemotherapy regimens and, thus, seeks to begin to fill the large void in clinical knowledge regarding the effectiveness of chemotherapy in elderly Medicare patients.